Logo UHasselt



Press and blog

Logo UHasselt Universiteit Hasselt - Knowledge in action


Findings from 1 year of CoMix in England

In our latest blogpost, we delve into the findings from the first year of CoMix study data in England. 

What is CoMix? 

The CoMix study is a groundbreaking social survey running across Europe, in which participants are asked to record the number of people they have come into contact with in the previous 24 hours. Participants are asked who they came into contact with (e.g. the age-group of the other person/s) and where the contact took place (e.g. whether at home or work, indoors or outdoors). They also answer questions about their attitude towards COVID-19. 

In England, the study has been running since March 2020 and almost 20,000 people have taken part. The individuals participating in the survey are representative of the English population as a whole in terms of age and gender. 

Between March 2020 and March 2021, England experienced 3 national lockdowns, interspersed with smaller, more localised restrictions. EpiPose researchers from LSHTM and UHasselt have taken a deeper dive into the first years’ worth of data to discover the ways in which people’s behaviour and movements have changed over the course of the pandemic. 

They compared the results from CoMix with data collected from earlier surveys which measured social contacts before the onset of COVID-19. 

The impact of restrictions 

As might be expected, government restrictions had a significant impact on the number of contacts people maintained; these reduced significantly during each national lockdown. Under full lockdown conditions where schools remained closed, working-age(18-59 years) people’s social contacts dropped to an average of 2.39. This is down from an average of 11.7 before the pandemic. 

However, lockdowns were not the only factor at play. Government restrictions were relaxed in Summer 2020. Although in-person contacts increased during this phase, people’s social contacts did not return to pre-pandemic levels. Among working-age adults, the average number of reported contacts during periods of relaxation was 4.93. Around half the number reported before the onset of COVID-19. 

Even during the Summer 2020 relaxation, social mixing did not really begin to increase until the government introduced incentives encouraging people to eat out at bars and restaurants. 

Individual differences

Some groups were likely to have more social contacts than others. For example, younger people - and especially those of school-age - typically had a greater number of contacts than older age-groups. Naturally, social contacts were tied to school reopenings. Young people’s contacts declined steeply when schools were closed, though this data should be interpreted with caution as contacts were reported by a parent rather than the child. People who worked part-time tended to have more social contacts than those in full-time work or the self-employed. Income did not appear to be a factor in determining how many social contacts an individual maintained. 

People’s perception of risk also influenced their mixing patterns. People who were less worried about becoming seriously ill from COVID-19 tended to have more social contacts than those who perceived the virus as a bigger threat. Similarly, those who fell into “high-risk” categories typically reported fewer social contacts than those who did not.  

How to interpret these results? 

Like any study of this kind, CoMix is not without limitations. People were recruited to the survey through the internet (e.g. email and social media), which not everyone has access to. Additionally, since adult participants self-report their social contacts and children’s contacts are reported by a parent, there is a chance that they may not recall every contact they had. 

Taking these points into consideration, the data collected in the survey still provides us with vital insights into people’s behaviour over the course of the pandemic. The findings can help policy makers think about the interventions which are likely to be successful in reducing contacts - and therefore curbing the spread of COVID-19 - as well as other factors which may influence people’s decisions. 

The study is available to read in full here.

***This research is a pre-print and has not been peer-reviewed***

Investigating the impact of lockdown and exit strategies: findings from Belgium

In this blogpost, we hear about how EpiPose researchers from the Universities of Hasselt and Antwerp used data from the CoMix study to investigate the impact of lockdown and the relaxation measures in Belgium. 

To curb the spread of COVID-19, Belgium entered a lockdown March 2020 which saw the closure of schools, universities, most workplaces and non-essential shops, and banned public gatherings. From early May, the Belgian government began to gradually lift restrictions. This was before vaccines became readily available. Few people were immune to the virus and the potential for another outbreak was of pressing concern.

What did they do? 

The team used a mathematical model to investigate the different phases of Belgium’s exit strategy. Assessing each phase for the potential impact on rates of COVID-19 transmission. 

  • Phase 1 (from 4th May): the team modelled for increased mixing as people began to return to work. 
  • Phase 2 (from 11th May): the team’s model predicted more social contact between children, as schools began to reopen. 
  • Phase 3 (from 8th June): the team modelled for more opportunities for in-person interactions to account for leisure and other activities reopening. 


The researchers needed to understand how patterns of social interaction might change over the course of the exit strategy - and the effect this could have on the spread of the virus. To generate realistic scenarios, they compared data on the number of people an individual came in close contact with (their ‘social contacts’) before the pandemic with information from during lockdown. They later cross-checked their findings with data collected from the CoMix study. CoMix is a survey which has been running throughout the pandemic and asks participants to record the number of people they came into close contact with in the previous 24 hours. Comparing their existing findings with the data from CoMix, the researchers got the same results. 

What did they find?

The model developed by the team was able to accurately describe the first wave of the pandemic, and initial exit strategy. It showed that reopening leisure activities would lead to the biggest increase in COVID-19 cases, followed by reopening workplaces. Reopening schools was less likely than measures like opening bars or restaurants to lead to a sharp rise in infections. 

Although they projected that cases would increase when Belgium exited lockdown, they found that this was not likely to lead to the wave of serious infections - and hospitalisations - as what was observed in the early stages of the pandemic. This could be due to people changing their behaviour when in close contact with others, for example by wearing masks. Or it could be the result of weather changes or other factors. People’s behaviour is constantly changing as the epidemic progresses, and it is important to carry out more research in this field to allow researchers to predict future trends. 

The team also found that tracking and isolating suspected COVID-19 cases was an effective way to curb the virus’ spread, whilst still allowing some social activities to take place. This demonstrates the vital importance of testing and tracing cases in keeping outbreaks of COVID-19 under control. 

The full study is available to read here.

Modelling the UK Government's 'Roadmap' out of lockdown in England

In Spring 2021, the UK Government announced a gradual easing of COVID-19 restrictions in England, known as the roadmap out of lockdown. 

In this blogpost, we learn how members of the EpiPose team are measuring the potential impact of the roadmap on SARS-CoV-2 transmission.

In late May 2021, the UK government relaxed several restrictions implemented in England as a result of the COVID-19 epidemic (step 3 of the roadmap). These relaxations include: reopening venues such as cinemas and indoor hospitality; lifting most restrictions on socialising outdoors; and allowing people to meet indoors in small groups. At the end of June 2021, the UK Government hopes to lift all legal limits on social contact (step 4 of the roadmap). 

These changes increase the potential for the SARS-CoV-2 virus to spread. However, the strong rollout of the country’s vaccination programme is a cause for optimism. So, what impact will the roadmap have on rates of transmission? 

On the 5th of May 2021, EpiPose researchers at the London School of Hygiene and Tropical Medicine presented a report addressing this question to the Scientific Pandemic Influenza Group on Modelling (SPI-M). SPI-M reports to the Scientific Advisory Group for Emergencies (SAGE), who provide scientific and technical advice to UK government decision makers during emergencies. The evidence presented by three modelling groups to SPI-M and SAGE is available on the UK government website.

What they did 

The team at LSHTM have developed a model that can project potential future transmission dynamics of SARS-CoV-2. The model used various data sources relevant to England. Sources include Google Community Mobility data and CoMix contact survey data, which are used to estimate the way that people mix over time in different settings (e.g. shops, pharmacies and train stations). Public health data on current vaccination coverage and future vaccination rates was used to estimate the number of people who would have additional protection provided by vaccines in the weeks ahead. 

The effect of relaxing restrictions will depend heavily on people’s behaviour. To get a comprehensive picture, the team projected several different scenarios about behavioural responses to steps 3 and 4 of the roadmap. They made projections for low, medium and high levels of in-person interaction and considered scenarios in which immunity (from vaccination or previous exposure to the virus) both did and did not decrease over time. 

Importantly, the team also looked at what might happen to SARS-CoV-2 transmission if new variants of concern (so called VOCs), which were more infectious or were able to evade immunity, were to emerge. 

What they found

Assuming no new variants of concern emerge, the EpiPose modellers predict that the easing of restrictions could push the effective reproduction number* above 1, leading to another wave of infections, hospitalisations and deaths during the summer months of 2021. The severity of this wave will depend on several factors, including: 

  • How effective vaccines are at protecting people from the virus 
  • How long people who have been vaccinated - or were infected in the past - remain immune 
  • The extent to which people change their behaviour as a result of the rules changing 
  • Whether ‘additional relaxation’ measures (such as lifting the requirement to wear masks) are introduced. 
  • Other factors, such as the weather, which may also impact transmission rates

The emergence of new variants of concern could paint a more worrying picture, according to the team’s projections, and would likely lead to large increases in hospitalisations and deaths. In particular, the emergence of a variant which was able to evade immune protection (developed after being infected by the virus in the past or from vaccinations) may prolong the epidemic. 

It is important to remember that these projections are made assuming that no counter-measures are implemented to contain the spread of the virus if transmission were to increase. 

How to interpret the results 

While these projections have been made using the best available data, the impact of lifting restrictions is difficult to predict with any certainty. The course of the epidemic will depend on how people’s behaviours change in response to relaxing restrictions, awareness of the epidemic itself, as well as other factors such as vaccination rates and vaccine effectiveness. 

The full report is available here

*To find out more about the reproduction number and other commonly-used COVID-19 terms, check out our ‘short guide to COVID-19 terminology’ on the blog!

Infectieradar: join the fight against COVID-19

In this blogpost, we explain what the Infectieradar platform is all about - and why your participation is crucial to its success! 

What is Infectieradar? 

Infectieradar is a web-based platform which collects data from people living in Belgium and the Netherlands about their current status of health and asks them to record any COVID-19 symptoms they may be experiencing. Alongside health-related information, the survey asks people about their behaviours and attitudes. 

Infectieradar is part of a larger network called Influenzanet, a European partnership between various universities and government health agencies. Influenzanet was established during the 2009 H1N1 Pandemic to collect information directly from the general public about symptoms of flu and has grown steadily since then. It is now active in 12 countries across Europe. 

Why are platforms like Infectieradar important? 

Infectieradar collects information directly from the population, which enables researchers to detect changes in COVID-19 infections at the earliest opportunity. Using web-based platforms like Infectieradar can help researchers and health authorities catch cases which might otherwise have been missed, as existing research has found. For example, we know that not everyone will make an appointment with their doctor when they begin to experience flu-like symptoms, but they may be willing to complete a very short survey. Through collecting information about their symptoms - alongside factors like whether they have spoken to a healthcare professional or are taking any medication - the platform can also alert people to the need to receive medical attention

Asking about behaviours and attitudes provides researchers with vital insights into how the pandemic is unfolding. As recent events have shown, individual behaviour amplified at population level can make a huge difference in being able to contain a health emergency. The information generated by platforms like Infectieradar is crucial not only now whilst we’re in the midst of a pandemic, but also in the longer-term. Proof of this can be seen with Influenzanet, which has been collecting high-quality information on influenza for over a decade. Because the platform was already active prior to the outbreak of COVID-19, it could be readily adapted to respond to this new health emergency. 

These platforms are only as good as the data that goes into them, which is why your involvement is invaluable. The more people that sign up, the better the data we’ll be able to capture and the quicker we’ll be able to respond to the virus. 

How can I join? 

If you live in Belgium, you can sign up to Infectieradar here. If you live in the Netherlands, you can join the platform here

What will I need to do? 

Each week, you’ll be asked to complete a short online survey about your current state of health. This should take no longer than a few minutes to complete. 

You can find out more here

What if I don’t live in Belgium or the Netherlands? 

Several countries across Europe have their own platforms which operate similarly to Infectieradar. You can find a list of these by visiting the Influenzanet website

We wish to offer a huge thanks to everyone who has joined these platforms to date, and to all those who continue to sign up. With your invaluable input, we’re able to monitor the evolution of the pandemic quicker than before, thereby curbing the spread and helping stop the virus in its tracks.

What impact do different local and national measures have on controlling COVID-19 spread? Findings from the CoMix study in England

In this latest blogpost, we consider how effective different interventions introduced in England over Summer and Autumn 2020 were on controlling the spread of coronavirus. During this time period, England moved out of a national lockdown, and moved to smaller and localised measures.

The interventions 

Areas experiencing higher rates of COVID19 were placed into localised lockdowns, rather than the nationwide closures seen before. The rules varied in different locations, but included: travel bans, non-essential venue closures, and restrictions on meeting indoors.

In mid-September, local lockdowns were supplemented by a number of nationwide restrictions, including: 

  • The ‘rule of six’: which limited the number of people who could socialise together
  • 10pm closures of bars and restaurants 
  • Encouragement to work from home, where possible.

In October, localised lockdowns were replaced by a three-Tier system where areas were classified; from medium (Tier 1) to high (Tier 2) to very high (Tier 3). The higher the Tier, the greater the restrictions

The aim of these various restrictions was to reduce people’s social contacts, thereby limiting the potential for coronavirus to spread. The more people an individual comes into contact with, the more likely they are to become exposed to the virus and pass it on to others. 

The study

To investigate the impact of these different rules, or ‘interventions’, on controlling the spread of the virus, members of the EpiPose team at the London School of Hygiene & Tropical Medicine used data collected from the CoMix study. They compared the number of social contacts a person recorded in different settings (e.g. school, work or home) before and after the introduction of various restrictions. Additionally, they examined how people’s contacts changed when moving from Tiered restrictions into the country’s second national lockdown in November. 

Using CoMix data, the team was able to calculate: 1. The proportion of people whose contacts decreased after restrictions were introduced and 2. The change in the average number of setting-specific social contacts someone had before and after measures were implemented. 

They found that participants were more likely to reduce their contacts after measures like the ‘rule of six’  compared to others like the early closure of bars and restaurants. And although the encouragement to work from home did see some people reduce their number of in-person interactions, almost two-thirds of people maintained the same number of social contacts before and after this recommendation was introduced. 

The introduction of localised lockdowns appeared to reduce people’s social contacts by a greater amount than under the Tier system. Though it is worth remembering that when the Tier structure came into place, other measures were already in force. There was little change in the average number of contacts people had before and after entering Tiers 1 and 2. Moving into Tier 3 appeared to slightly reduce the average number of contacts a person kept, however the number of study participants who were subject to Tier 3 restrictions was quite small, making this finding difficult to generalise. 

The effectiveness of the November lockdown on reducing an individual’s social contacts depended on which Tier they entered from. Those who moved into lockdown from Tier 1 reported the biggest reduction in their average number of contacts. Those moving from Tiers 2 or 3 into lockdown reported fewer changes in their contact patterns, potentially because they were already minimising the number of people they came into contact with. 

How should we interpret these findings?

The impact of these measures on reducing a person’s social contacts are decidedly less pronounced than under the first national lockdown in March. Yet, virtually all were successful in limiting the number of contacts people had. Albeit to varying degrees. 

It’s also important to consider other factors at play. Many people may have already been limiting their number of in-person interactions prior to the official measures coming into force. For instance, by working from home or limiting contact with friends or family. As some of these restrictions occurred at the same time, and because policies changed rapidly over the course of a few short months, it is difficult to fully attribute changes in people's behaviour to the introduction of specific restrictions. 

Taking these factors into account, the findings of this study still provide policymakers with valuable information in considering what kinds of interventions to introduce in response to COVID19.

The full article this blogpost is based on is available to read here: The impact of local and national restrictions in response to COVID-19 on social contacts in England: a longitudinal natural experiment


A short guide to COVID-19 terminology

Reading the coverage of the coronavirus pandemic over the past 12 months, you will likely have noticed many terms which simply weren’t in most of our vocabularies in 2019. 

But what do some of these words and phrases mean? And how are they useful for interpreting the course of the pandemic? To kick off our EpiPose blog post series, we explain what some of these terms mean and why they’re used to discuss COVID-19 and within the EpiPose project.  

Epidemiology: Epidemiology is the study of how, why and where diseases (or other health-related states or events) develop in a given population. Epidemiologists track the patterns, causes and risk-factors associated with a given disease or health state with the aim of  understanding how populations can be healthier. To date, this field of study has proved crucial in the response to COVID-19 and will continue to inform how policy makers respond to the pandemic. 

Epidemiological parameters: These are the quantities epidemiologists use to describe a disease (and for an infectious disease, how it spreads), and include many of the items listed below; 

Incubation period: This is the time between someone becoming infected with the virus and when they begin to develop symptoms. 

Asymptomatic: If a person is asymptomatic, it means that they are infected with the virus but do not develop symptoms. 

Presymptomatic: A person who is presymptomatic has been infected with the virus and is not currently showing symptoms. However, they will go on to develop symptoms later. There is a danger that people who are asymptomatic and presymptomatic have the potential to infect a larger number of people because they aren’t aware they’re infected.

R number: The R (or Reproduction) number is the average number of people a person who has been infected will pass a virus onto, over the entire course of that infection. For example, an R value of 5 would mean that once infected, someone would pass the virus to an average of 5 other people. The R number can be reduced if intervention measures (such as school closures or travel restrictions) are put in place to stop a virus from spreading. 

K number: While the R number calculates the average number of people someone with the SARS-CoV-2 virus will transmit the virus to, not everyone will transmit to the same number of people. The K number estimates the variation in how many people each person with the virus infects. 

Case fatality risk(CFR): The proportion of confirmed COVID-19 cases which prove fatal. This can be calculated by dividing the number of confirmed COVID-19 deaths from the number of confirmed cases. 

Infection fatality risk (IFR): This is the proportion of all COVID-19 cases (both diagnosed and not diagnosed) which prove deadly. It’s calculated by dividing the total number of infections by the number of deaths and is a more difficult figure to estimate than the CFR, as it involves identifying COVID-19 cases which may have been missed from official figures.

Vaccine efficacy: This is the % reduction in disease in vaccinated people compared to unvaccinated people in a controlled environment. For instance, in a clinical trial setting. 

Vaccine effectiveness: This is the % reduction in disease in vaccinated people compared to unvaccinated people in real-world settings. 

And one final question:

Why do we see both COVID-19 and SARS-CoV-2 being used in communication material about the coronavirus? 

SARS-CoV-2: stands for severe acute respiratory syndrome coronavirus 2 and is the virus which causes the coronavirus disease. COVID-19: refers to the disease which can be contracted from being exposed to the virus. The World Health Organisation have provided some useful information about why we see both terms being used to discuss coronavirus. 

The impact of reopening schools on SARS-CoV-2 transmission in England: evidence from the CoMix study 

Fully reopening schools could push the reproduction number (R) of SARS-CoV-2 in England above 1.0, potentially putting an end to the decline in new cases, suggests a new pre-print. The modelling study, not yet peer-reviewed, was conducted by members of the EpiPose team and colleagues at the London School of Hygiene & Tropical Medicine (LSHTM).

Schools present more opportunities for the virus to be transmitted so are an important consideration when looking at the spread of COVID-19. In January 2021, the Government in England announced the closure of primary and secondary schools as part of the country’s third national lockdown. However, there are concerns about the potentially damaging impact closures may have on students’ academic development and general wellbeing. To date, the evidence on how effective school closures have been in curbing the spread of the virus remains unclear. 

Previous studies suggest children may be less likely than adults to be infected upon exposure to the virus, and may differ in their infectiousness too. To account for these differences, the team applied various scenarios of how susceptible and infectious school-aged children were compared with adults.

The study used data collected from the CoMix study, measuring children and adults’ social contacts during the November 2020 and January 2021 lockdowns, exploring how this behaviour changed with the closure of schools in the latter. This was combined with different estimates of children’s susceptibility (how likely they are to be infected upon contact with an infected individual) and infectiousness (how easily they infect others). 

Using official estimates of the current R number, the team estimated the possible increase in R upon opening schools. Across all estimates of children’s infectiousness and susceptibility, the effect of fully reopening schools saw R increase from an assumed baseline of 0.8 to between 1.1-1.5. Partial school reopening - primary or secondary schools only - resulted in lower increases and, dependent on children’s susceptibility and infectiousness, could see R increase to between 0.9-1.2. 

These findings offer crucial insights to support decisions about whether, and how, schools should reopen. An R number greater than 1.0 - as some of the estimates in this research predict - signals that the epidemic would begin to grow and that cases are likely to increase. 

However, it is important to remember that these estimates will shift based on changes in other factors that affect R, such as changes in relative susceptibility due to variation in ongoing infection rates between age-groups and the emergence of different variants of the virus. Furthermore, our understanding of children’s susceptibility and infectiousness as compared to adults is still evolving and more precise estimates are needed. The situation is likely to change further as we learn more about the new variants of COVID19 and their patterns of transmission. 

The study is available to read in full here:

***This research is a pre-print and has not been peer-reviewed***  

The original news story about this study is available at: